**Growth on Metallo-Supramolecular Coordination Polyelectrolyte (MEPE) Stimulates Osteogenic Di**ff**erentiation of Human Osteosarcoma Cells (MG63) and Human Bone Marrow Derived Mesenchymal Stem Cells**

#### **Janina Belka 1, Joachim Nickel 2,3,\* and Dirk G. Kurth 1,\***


Received: 24 May 2019; Accepted: 26 June 2019; Published: 27 June 2019

**Abstract:** Background: Culturing of cells is typically performed on standard tissue culture plates generating growth conditions, which in general do not reflect the native three-dimensional cellular environment. Recent investigations provide insights in parameters, which strongly affect the general cellular behavior triggering essential processes such as cell differentiation. The physical properties of the used material, such as stiffness, roughness, or topology, as well as the chemical composition of the cell-surface interface are shown to play a key role in the initiation of particular cellular responses. Methods: We extended our previous research, which identified thin films of metallo-supramolecular coordination polyelectrolytes (MEPEs) as substrate to trigger the differentiation of muscular precursor cells. Results: Here, we show that the same MEPEs similarly stimulate the osteogenic differentiation of pre-osteoblasts. Remarkably, MEPE modified surfaces also trigger the differentiation of primary bone derived mesenchymal stem cells (BMSCs) towards the osteogenic lineage. Conclusion: This result leads to the conclusion that these surfaces individually support the specification of cell differentiation toward lineages that correspond to the natural commitment of the particular cell types. We, therefore, propose that Fe-MEPEs may be used as scaffold for the treatment of defects at least in muscular or bone tissue.

**Keywords:** cell differentiation; metallo-supramolecular polymer; interface; iron metabolism

#### **1. Introduction**

In a previous study, we investigated the influence of Fe-MEPE modified surfaces on the differentiation of pre-myoblastic C2C12 cells in standard cell culture medium [1]. C2C12 cells differentiate towards myotubes after being seeded onto Fe-MEPE modified surfaces without the need to add agents that promote differentiation. In case of C2C12 cells, if grown on standard cell culture plates, myogenic differentiation is typically initiated by serum starvation [2]. To verify the hypothesis that culturing cells on Fe-MEPE modified surfaces might in general stimulate cell differentiation towards pre-determined differentiation lineages, we extended our investigations on differentiation of the human osteosarcoma cell line MG63 and that of primary human bone marrow derived mesenchymal stem cells (BMSCs).

Using standard cell culture plates differentiation of MG63 cells generally requires addition of special agents such as ascorbic acid and dexamethasone [3,4]. Standard protocols for the osteogenic differentiation of primary pluripotent BMSCs also utilize dexamethasone and ascorbic acid, but may also require addition of bone morphogentic protein (BMP)2 and β-glycerophosphate [5]. Dexamethasone primarily promotes cell proliferation whereas ascorbic acid induces expression of alkaline phosphatase (ALP) and osteocalcin [6,7]. Despite of their osseous origin, BMSCs can be differentiated towards different lineages in vitro such as osteoblasts, chondrocytes, astrocytes, neurons, skeletal, and cardiac muscles [8–11]. Even though these cells are pluripotent, they are committed to differentiate towards bone cells.

As a general sign of differentiation, cellular parameters such as cell growth and mitochondrial activities are assessed since high mitochondrial activities at low cell count indicates cell differentiation. The osteogenic differentiation process is characterized by several steps. At first, the cells adhere and proliferate. Subsequently, the extracellular matrix is formed, which is finally followed by mineralization. Collin et al. considered clustering of cells before matrix formation as a separate step [12].

During the formation of extracellular matrix the osteoblasts produce collagen type I, osteocalcin, osteopontin, and alkaline phosphatase (ALP), which drastically increases its activity at the beginning of bone matrix mineralization [13]. The high activity of ALP leads to an increased release of phosphate, which forms the mineral part of the bone with free calcium ions [14]. At onset of osteogenic differentiation, ALP is up-regulated, whereas osteocalcin is primarily expressed at a later stage upon mineralization. According to Collin et al., ALP is expressed at the beginning and at the end of these differentiation processes [12]. Owen et al. assume that ALP increases linearly during differentiation and drastically increases before mineralization [15]. With the beginning of mineralization, osteocalcin reaches its maximum of expression [16]. According to Bronckers et al. increased osteocalcin expression occurs in osteoblasts and osteocytes [17].

However, it is known that primary cells do not differentiate synchronously in vitro [18]. Thus, in a pool of cells, both ALP and osteocalcin expression may be detected at the same time. Additionally, cell differentiation is dependent on the site of cell collection, the number of residing cells, the gender of the donor, and on cell extraction and purification methods [19]. However, a detailed understanding of the orchestrated interaction and regulation of operating signaling pathways and regulating factors involved in these differentiation processes is not available yet.

Several studies on the effects of metal ions on cellular systems have been published in the past [20]. Concerning osseointegration metal ions were shown to play an important role in the processes of angiogenesis, osteogenesis, and mineralization of bone tissue. Above all, Cu(II)- and Co(II)-ions stimulate the secretion of certain growth factors, e.g., vascular endothelial growth factor (VEGF) via hypoxia inducible factor (HIF)-1, subsequently stimulating proliferation of endothelial cells and, thus, the formation of blood vessels [21]. The cations Zn(II) and Sr(II) instead stimulate osteoblast proliferation while inhibiting osteoclast activity [22,23]. These ions thus have a stimulating or anabolic effect on bone homeostasis. The cations Ca(II) and Mg(II) stimulate osteoblast proliferation. The latter also stimulates cell adhesion via binding of integrins to extracellular matrix proteins [24].

In this study, we investigated the effect of Fe-MEPE modified surfaces on the proliferation and differentiation of MG63 cells and BMSCs. As in our previous publication, the metallo-supramolecular coordination polyelectrolyte (MEPE) based on the ditopic ligand 1,4-bis (2,2 :6 ,2 -terpyridin-4 -yl) benzene and Fe(II)-ions was used [1]. As several methods can be employed in order to deposit the positively charged Fe-MEPE on surfaces, we used the layer-by-layer deposition method and dip-coating from solution [25–30]. Due to the dynamic nature of the interaction of terpyridines and metal ions, such as Fe(II), Ni(II), or Co(II), MEPEs may represent ideal metal ion release systems. The released ions may stimulate specific cellular responses in target cells, which adhere to the modified surface.

#### **2. Materials and Methods**

#### *2.1. Surface Modification*

#### 2.1.1. Layer by Layer

Layer-by-layer coating was carried out using the method described earlier [27]. Circular white borosilicate glass slides with a diameter of 18 mm were washed with ethanol and dried with compressed air before. The polyelectrolytes were dissolved in aqueous 0.1 M sodium acetate solution to improve the layer thickness by weakening the intra-molecular electrical repulsion of the polymer chains [31]. First, a layer of polyethyleneimine (PEI, Fluka) was deposited on the glass slide using a coating solution with a concentration of 10−<sup>2</sup> M. For the formation of the second layer, a 1 <sup>×</sup> 10−<sup>3</sup> M poly-(styrene sulfonate) (PSS) solution was used. The final MEPE-layer was adsorbed using a 2.1 <sup>×</sup> <sup>10</sup>−<sup>3</sup> M Fe-MEPE solution. An incubation time of 4 minutes was chosen. Between the coating steps, the substrates were rinsed with ultrapure water. After the application of the last layer, the samples were dried with compressed air [27,29]. The resulting LbL sample consisted of a PEI, a PSS, and a final Fe-MEPE Layer.

#### 2.1.2. Dip Coating

Glass slides were washed with deionized water and finally with absolute ethanol. Fe-MEPE was dissolved in absolute ethanol with a concentration of 8.8 <sup>×</sup> <sup>10</sup>−<sup>3</sup> M [28]. Dip-coating was performed at ambient temperature by pulling the immersed glass slides at constant speed of 10 mm/min (Dip10) or 50 mm/min (Dip50) out of the coating solution. The slides were air-dried at ambient temperature.

#### *2.2. Cell Culture and Biological Activity Testing*

The MG63 (ATCC Number CRL-1427) and hMSCs (Donor 46: male, 73 years, Donor 54: female, 61 years, Donor 56: female, 75 years) cells were cultured in Dulbecco's Modified Eagle's Medium containing 10% FCS and 1% Penicilin/Streptomycin. For all experiments, 10,000 cells/cm2 were seeded onto the different substrates. Analysis of cell viability and -proliferation of the MG63 was performed 3 and 5 days after seeding. The hMSCs were analyzed after 2, 5, 7, and 9 days. To differentiate the two parameters, proliferation and metabolic activity, the cells were incubated with Dulbecco's Modified Eagle's Medium containing WST-1 (Roche). The metabolic activity was measured with an ELISA microplate reader (TECAN infinite M200) at 450 nm. After measurement, the cells were trypsinized and the cell number determined using a Neubauer counting chamber.

#### *2.3. RNA Isolation and cDNA Synthesis*

For RNA isolation, MG63 cells and hMSCs were trypsinized, subsequently washed with PBS, and centrifuged. RNA from the pelleted cells was isolated using the RNeasy Micro Kit (Qiagen, D-40724, Germany) according to the manufacturer's instructions. RNA concentration and purity was determined by spectralphotometry at wavelengths of 260 nm and 280 nm, respectively. 500 ng of each RNA sample was reversely transcribed using the iScriptTM cDNA Syntheses Kit (BioRad, Hercules, CA, USA), according to the manufacturer´s recommendations.

#### *2.4. Quantitative RT-PCR*

Differentially expressed genes were quantified by qRT-PCR. The used primers are: hGAPDHsense, 5 -TGACGCTGGGGCTGGCATTG-3 and hGAPDHantisense, 5 -GCTCTTGCTGGGGCTGGTGG-3 , ALPsense, 5 -CTTGACCTCCTCGGAAGACACTC-3 and ALPantisense 5 -GCCTGGTAGTTGTTGTG AGCATAG-3 , Osteocalcinsense 5 -TTGGACACAAAGGCTGCAC-3 , and Osteocalcinantisense, 5 -CTCACACTCCTCGCCCTATT-3 .

PCR-reactions were carried out using a C1000TM Thermal Cycler including a CFX96 real-time monitoring system according to Belka et al. [1].

#### *2.5. Colorimetric Determination of Iron(II)*

The iron content of the cells was determined by a colorimetric assay based on ferrozine complexation. With this method it is possible to detect iron contents with a lower limit of 0.2 nmol. In brief, Fe(III) and Fe(II)-ions have at first to be removed from proteins and the cellular network. For that the cells were washed thoroughly and lysed before the iron could be removed quantitatively from iron-loaded proteins like ferritin or heme-proteins. Subsequently, all iron species were reduced since ferrozine only forms complexes with iron (II). Cell lysis and complex formation by ferrozine is described in further studies [1].

#### **3. Results**

#### *3.1. Preparation and Characterization of Modified Surfaces for Cell Growth*

It is well understood that the physico-chemical properties of an interface greatly affect interactions with cells [32]. Therefore, meticulous attention is paid to the preparation and characterization of modified surfaces. The surface modification and characterization follows the experimental protocol established in a previous study [1]. Two methods of surface modifications are employed for the current study. First, the layer by layer deposition is used to deposit the metallo-polymer on the surface [27,29,33]. The final films are composed of a polyethyleneimine primer, poly-(styrene sulfonate) (PSS), and finally Fe-MEPE. Second, Fe-MEPE is directly applied from solution onto the substrate by dip-coating [28]. The final layer thickness amounts to 25 ± 1 nm (FeDIp10) and 43 ± 2 nm (FeDip50). The roughness of the modified surfaces is in the range of 3–7 nm [1]. The mean values of the contact angles are 69 ± 2◦ for the Dip10 and 67 ± 2◦ for the Dip50 substrates, indicating that both substrates are hydrophilic (data not shown) [34]. The contact angle of the LbL surface is 57 ± 2◦ [1].

#### *3.2. Cell-Substrate Interactions*

First, we investigated the cell number and the metabolic activity of the cell line MG63 to assess the cytotoxicity of Fe-MEPE in relation to the TCPs-reference (tissue culture polystyrene) surface. In Figure 1, cell number (A) and activities (B) of MG63 cells grown on the functionalized surface are shown for day 3 and 5.

**Figure 1.** (**A**) Cell number and (**B**) metabolic activities of the used MG63 cells was determined at day 3 and 5 for the particular surface modifications. The cell number and the metabolic activity were set in relation to cells grown on tissue culture polystyrene (TCPS) substrates (reference). The experiments were performed at least two times in triplicate. \* *p* < 0.5 (analyzed by one-way ANOVA with Tukey test).

As can be seen, application of Fe-MEPE modified surfaces had a significant influence on both, cell number and cell activity of the used osteosarcoma cells. At both days, the evaluated cell number

for the functionalized substrates was below those of the reference substrate. The activity of the FeDip10 modified surface was 82 ± 12%, which is within the non-cytotoxicity limit defined for the range of 81–100%. In the case of the FeDip50 layer, a significantly reduced proliferation at day 3 can be observed, thus classifying this surface as cytotoxic with a degree of 3 [35]. Relative to the reference substrate, the cell number for the FeDip50 modified surface dropped from 41 ± 12% at day 3 to 26 ± 7% at day 5. The relative cell number on the LbL-modified surface decreased similarly from day 3 to day 5. Here, the values decreased from originally 74 ± 14% to 25 ± 7%. The FeDip10 modified surfaces showed moderately reduced cell proliferation from day 3 to 5, with values ranging from 82 ± 12% to 66 ± 11% compared to the reference substrate.

Comparing the cell activity on the basis of the mitochondrial activity per cell Figure 1B shows that the cells stopped proliferation in favor of metabolic activity, reaching 378 ± 31% at day 5 for FeDip50 modified surfaces compared to cells grown on the TCP reference surface. Cells grown on LbL modified surfaces showed an increase in metabolic activity of 139 ± 2 5% and FeDip10 modified surfaces of 98 ± 21%, respectively. These high metabolic cell activities suggest that cells grown on Fe-MEPE modified surfaces might have been stimulated to differentiate most likely to the osteogenic lineage. To confirm this hypothesis, alizarin red staining was carried out in order to detect a potentially increased Ca(II)-ion storage in the cells. Figure 2 shows microscopy images of the alizarin red stained cells at day 3.

**Figure 2.** Light microscopy images of Alizarin red stained MG63 cells. (**A**) Cells grown on Reference substrate, (**B**) LbL, (**C**) FeDip10, and (**D**) FeDip50 at day 3.

The overall morphology and the cell number of cell grown on the reference substrates differed greatly as seen in the images taken from the individual surfaces. On the FeDip10 modified surfaces cells did not reach confluency, while cells grown on the reference substrate appeared confluent.

The low cell numbers are better visualized in Figure 3, showing Alizarin red stained cells at two different magnifications at day 5.

**Figure 3.** Light microscopy image**s** at different magnification of Alizarin red stained MG63 cells grown on (**A**,**B**) reference substrates, (**C**,**D**) FeDip10, and (**E**,**F**) on FeDip50 substrates at day 5.

Based on cell morphology, cluster formation correlating with Alizarin red staining indicates osteogenic differentiation on the FeDip50 modified surfaces already on day 3, which represents the second differentiation stage according to Collin et al. [12].

By detachment and reseeding after growth for three days on Fe-MEPE modified surfaces, the already-formed clusters were disrupted, and thus, the cells had to re-initiate themselves. Due to the very low relative proliferation rates of the MG63 cells on Fe-MEPE modified substrates (see Figure 1) only small cell deposits on dip-coated Fe-MEPE modified surfaces can be seen at day 5. However, cells significantly differing in the overall morphology can be observed (Figure 3F), which were also characterized by increased Ca(II)-ion deposition (Figure 2D). On all Fe-MEPE modified surfaces, the cells adhere strongly with extended cell protrusions. As expected and in agreement with previous experiments, MG63 cells differentiate according to their initial origin (osteogenic) on Fe-MEPE modified surfaces [1]. In order to determine whether the used Fe-MEPE modified surfaces support in general differentiation process of already committed cells, we expanded our experiments by including human primary cells, so-called bone marrow derived mesenchymal stem cells (BMSCs) obtained from the spongiosa of femur bones.

As in the previously reported experiments, BMSCs are seeded and grown for up to 9 days on LbL, FeDip10 and FeDip50 modified surfaces. As in Figures 1 and 2, cell count and cell activity of BMSCs derived from three donors which are grown on LbL, FeDip10 and FeDip50 modified surfaces as shown in Figures 4 and 5.

**Figure 4.** Cell growth of the used bone derived mesenchymal stem cells (BMSCs) from (**A**) Donor 46, (**B**) Donor 54, and (**C**) Donor 56 relative to that on the reference surface for the particular surface modifications at four different time points. All experiments were performed in triplicate. \* *p* < 0.5 (analyzed by one-way ANOVA with Tukey test).

In general, similar results were obtained regarding proliferation of cells obtained from donor 46 and 54 on the differently modified surfaces compared to the reference surface. As shown for MG63 cells above, we note here the most prominent reduction in cell growth of 56 ± 16% for cells (donor 46) grown on FeDip50 modified substrates compared to the reference substrate. For donor 54, a reduction to 60 ± 13% compared to the reference substrate is observed. A reduction of proliferation for cells grown on LBL-modified surfaces as shown for MG63 cells could not be observed. In contrast, cells of donor 56 appear to proliferate much more on LBL- and Dip10-modified surfaces, whereas cell growth on FeDip50-modified surfaces is comparable to that of the reference substrate. Cells of this donor thus seem to differentiate less than cells of donor 46 and 54.

In order to proof the above mentioned hypothesis, the cell activity is tested by WST-1 assays. On the one hand, the observed low cell numbers might be caused by cell death or, on the other hand, are the result of a proliferative stop, which typically accompanies differentiation processes. If the cells are simply not viable due to a lack of cell adhesion or to a release of cytotoxic substances from the

particular MEPE-modified surfaces, no increase in metabolic activity should be observed in these assays. As shown in Figure 5, the determined metabolic activity depicted as mitochondrial activity per cell, argue against cell death since values obtained from cells grown on the MEPE modified surfaces are throughout the duration of the experiment at least equal to those determined for cells grown on the reference substrates. For samples employing cells from donor 46 and 54 a clear trend towards an increased cell activity is apparent from day 5. At day 5, the relative activity of cells from donor 46, which are grown on FeDip50 modified substrates, are found to be 180 ± 67% compared to the reference surfaces. For cells derived from donor 54 this value rises to over 300% of that of the reference. In contrast, for cells grown on LbL-modified substrates the metabolic activities are not increased compared to the reference.

**Figure 5.** Relative Metabolic activities of the used hMSC cells from (**A**) Donor 46, (**B**) Donor 54, and (**C**) Donor 56 was determined at four different time points for the particular surface modifications in relation to cells grown on reference substrates. All experiments were performed in triplicate. \* *p* < 0.5 (analyzed by one-way ANOVA with Tukey test).

Thus, based on these observations, we assume that cells, which are grown on the dip-coated surfaces, tend to differentiate. Due to the osteogeneous origin of the used BMSCs, a differentiation towards osteogenic lineages is expected, which can be assessed by specific assays such as Alizarin red staining (as shown for MG63 cells, see Figure 3) or by determination of alkaline phosphatase (ALP) activity.

Figure 6 shows microscopy images of Alizarin red stained cells of donor 46 and 54 on the reference and FeDip50 surface at day 5.

**Figure 6.** Examples of light microscopy images of Alizarin red stained BMSCs grown on (**A**) reference (Donor 409 46), (**B**) FeDip50 (Donor 46), and (**C**) FeDip50 (Donor 54) at day 5.

It becomes obvious that the cell morphology on the reference has not changed. Importantly, no staining by Alizarin red is detectable indicating that these cells reside in an undifferentiated state, which supports the data obtained for the cell activity (see Figure 5). In clear contrast, cells grown FeDip50 modified surfaces exhibit a different spindle-like phenotype. The cells are assembled in form of clusters, which are positively stained by Alizarin red. The cells on the reference surface appear denser than those grown on dip coated surface. This observation is in agreement with the determined cell numbers shown in Figure 4. Based on these stainings we conclude that the cells grown on FeDip50 modified surfaces have reached after stage 2 5 days differentiation in accordance with Collin et al. [12]. Thus, altered cellular activity and cell morphology indicate differentiation of the used cells towards the osteogenic lineage as supported by the positive alizarin red staining.

#### *3.3. Iron Content*

In order to show whether potential differences in the intracellular iron content may depend on the particular surfaces, the cells are removed from the substrates, washed, lysed, and the iron ions released from the carriers such as ferritin and hem-proteins are detected by photometric analysis. The relative content of iron ions of cells grown on the particular Fe-MEPE surfaces compared to the reference surface are shown in Figure 7.

In agreement to the data concerning cell number and activity, cells from donor 46 and 54 grown on Fe-MEPE modified surfaces show, as expected, an increased intracellular iron ion content over a prolonged time of 9 days compared to the reference. On the other hand, cells from donor 56 show an initial increase in iron ion content on day 3; however, starting from day 5, no significant differences in iron ion content compared to the control could be detected.

The influence of certain metal ions on BMSCs has already been shown. For instance, Yoshizawa et al. detected a rapid proliferation of BMSCs and an increase in extracellular matrix (ECM) mineralization in the presence of Mg(II)-ions in vitro [36]. On the other side, transition metal ions such as Mn(II), Fe(III), Co(II), Ni(II), and Cu(II), showed cytotoxic effects on osteoblastic cell lines such as MG63 cells

already at a concentration of 0.1 mM [37]. Typically, osteogenic differentiation of BMSCs is initiated by addition of various agents such as dexamethasone, L-ascorbic acid-2-phosphates or ascorbic acid and β-glycerophosphates. As the aforementioned agents are absent in the experiments shown here, we assume that solely the Fe-MEPE modified surfaces trigger differentiation of the used BMSCs. However, the determination of the concentration of the active species acting on the cell in the interfacial region remains to be elucidated. As positive staining by Alizarin red already indicates differentiation of the investigated BMSCs towards osteogenic lineages this process is analyzed in more detail by investigating the expression profiles of characteristic marker genes such as ALP and Osteocalcin (OC) by qRT-PCR.

**Figure 7.** The bar diagrams represent the relative ferrous concentrations per cell for (**A**) donor 46, (**B**) donor 54, and (**C**) donor 56 grown on the indicated substrates. The values obtained for cells grown on TCPS substrates are set to 100% (the values for the references substrates also vary donor dependently. The absolute values for the reference substrates at day 3 are: donor 46 = 68.2 fMol/cell; donor 54 = 31 fMol/cell; donor 56 = 21.6 fMol/cell). All experiments were performed in triplicate. \* *p* < 0.5 (analyzed by one-way ANOVA with Tukey test).

#### *3.4. Quantification of ALP and OC Gene Expression*

BMSCs, when grown for 16 days in the aforementioned differentiation media, form a coherent network of ALP-positive cells, but when Co(II)-ions at a concentration of 40 μM to 100 μM are added, a decrease in the overall ALP activity is observed [38,39].

Figure 8 shows ALP and OC gene expression at day 3 and day 5 of the three donors.

**Figure 8.** The bar diagrams represent values for relative ALP (**A**,**C**,**E**) and OC (**B**,**D**,**F**) expression levels of cells grown on the indicated substrates of Donor 46 (**A**,**B**), Donor 54 (**C**,**D**), and Donor 56 (**E**,**F**). All values are normalized to GAPDH expression levels. Subsequently the levels of ALP and OC expression determined for cells grown on reference substrates are set to 1. All experiments were performed in triplicate.

In contrast to the clear results obtained from experiments regarding cell number and cell activity those addressing ALP and OC expression need to be described and discussed in more detail. Unexpectedly, ALP expression in cells of donor 46 grown on FeDip modified surfaces is in general

comparable to that of cells grown on the reference substrates. At day 5, the values for ALP activity decrease by half relative to the reference. For the LbL modified surface lower ALP expression is observed at day 3, but it rises almost to the reference value by day 5.

Since these findings are not in agreement with our data regarding cell number and activity, we also analyzed the expression of OC. Here, a significant increase in OC expression is observed for cells of donor 46 grown on FeDip10 and FeDip50 modified surfaces at day 3 if compared to the reference substrate. However, by day 5 OC expression in these cells declines strongly to values of approximately 10% of those of the reference. Instead, cells of donor 54 grown on the Fe-MEPE modified surfaces in general show higher ALP expression levels if compared to the controls. Astonishingly, cells of donor 56 show high expression of either ALP and OC but only if grown on FeDip50 modified surfaces at day 3, which is unexpected since data concerning cell number and activity identify cells of this donor to be the least differentiating ones.

Taken together, the results of these experiments suggest that Fe-MEPE-triggered differentiation strongly depends on the individual surface modification and, as expected, also on the individual donor thus preventing an exact determination of particular differentiation stages. Alkaline phosphatase, as already mentioned in the introduction, is primarily expressed upon osteogenic differentiation in two separated time frames. A first peak is observed at the beginning of osteogenic differentiation and a second at the beginning of mineralization. As shown by individual microscopy images, only sub-populations of cells can be stained by alizarin red implicating that cell differentiation, even in one specified sample, is non-homogenous. Thus, the data obtained from qRT-PCR experiments, representing the average ALP or OC expression levels of these non-homogenously differentiated cells, may not indicate a general differentiation stage.

However, it can be assumed that differentiation of at least some of the cells of donor 46 reached the second differentiation step already at day 3, which is confirmed by alizarin red staining (see Figure 8) but a comprehensive progress in differentiation till day 5 cannot be detected

Collin et al. observed an increase in osteocalcin expression by 30-fold during mineralization [12]. This considerable increase in osteocalcin gene expression might be achieved for BMSCs of donor 56 simply grown for 3 days on FeDip50-modified surfaces. Birgani et al. reported no increased ALP or osteocalcin gene expression until day 14 after metal ion addition. Moreover, Birgani et al. suggests that the ALP activities of the different donors can vary [40]. These different activities as a function of the BMSC donor have already been first described by Barbara et al already in 2004 but meanwhile also by many others [41–45]. The variety of gene expression between the different donors has also been observed in our study. But the duration of ALP and OC gene expression was reduced from 14 to only 3 days.

However, the results presented here indicate that osteogenic differentiation is initiated by Fe-MEPE modified surfaces and can be detected at day 3. In this study we detect ALP expressing cells already at day 5. Thus, the use of Fe-MEPE modified surfaces seems to trigger the osteogenic differentiation processes more rapidly.

#### **4. Conclusions**

The current investigation employing MG63 cells grown on Fe-MEPE modified substrates suggest initiation of osteogenic differentiation by both, high cell activity and altered morphology of the cells and/or cluster formation. Remarkably, LbL and FeDip50 modified surfaces show strongest effects on cell count and cell activity, which becomes visible already at day 3. Considering morphology, cells grown on LbL modified surfaces appear morphological similar and do not form cell clusters like those grown on the reference substrate. Based on these observations, MG63 cells seem to best differentiate to osteoplastic lineages if grown on these substrates, which is further supported by the occurrence of Alizarin red stained clusters (Figure 2).

Concerning BMSCs, our findings suggest that Fe-MEPE modified surfaces also stimulate osteogenic differentiation in these cells. Similar to MG63 cells, the Fe-MEPE modified surfaces suppress proliferation and promote differentiation of the used BMSCs to a variable extent which is dependent on the individual donor. Cells derived from two donors (46/54) differentiate better than cells of the third donor (56), which is supported by staining with Alizarin red. Clearly, BMSCs are stimulated for osteogenic differentiation, which appears donor and substrate specific. However, cell differentiation occurs non-coherently thus reflecting various differentiation stages already in one particular sample.

The effect of the modified surfaces on osteogenic differentiation can also be detected on mRNA level addressing well-known osteogenic marker genes, such as ALP and OC, although the data do not show the expected coherency. Nevertheless, all Fe-MEPE modified surfaces, which are investigated here, influence the osteogenic differentiation capacity of BMSCs without addition of agents inducing osteogenic differentiation. This remarkable result leads to the question of the underlying mechanism. It is well established, that metal ions affect cell differentiation and, therefore, most likely also a release of iron ions into the interfacial contact region between substrate and cells initiate cellular processes towards differentiation. The osteogenic properties of iron ions in context of BMSC differentiation has also been reported for iron oxide nanoparticles by Wang et al. [46,47]. Thus, our results indicate that Fe-MEPE functionalized surfaces may serve as innovative scaffolds for the treatment of bone defects.

**Author Contributions:** Conceptualization, J.N. and D.G.K.; methodology, J.N. and J.B.; validation, J.B.; investigation, J.B.; writing—original draft preparation, J.B.; writing—review and editing, J.N. and D.G.K.; supervision, J.N.; project J.N. and D.G.K.

**Funding:** This publication was funded by the German Research Foundation (DFG) and the University of Wuerzburg in the funding programme Open Access Publishing.

**Conflicts of Interest:** The authors declare no conflict of interest

#### **References**


© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

## *Article* **Biotechnological Preparation of Gelatines from Chicken Feet**

#### **Pavel Mokrejš 1,\*, Petr Mrázek 1, Robert Gál <sup>2</sup> and Jana Pavlaˇcková <sup>3</sup>**


Received: 15 May 2019; Accepted: 14 June 2019; Published: 18 June 2019

**Abstract:** In the European Union (EU), about five tons of poultry by-product tissues are produced every year. Due to their high collagen content, they represent a significant raw material source for gelatine production. The aim of the paper was the biotechnological preparation of gelatine from chicken feet. The influence of selected process factors on the gelatine yield, gel strength, viscosity, and ash of gelatine was observed; a two-level factor design of experiments with three variable process factors (enzyme addition, enzyme treatment time, and gelatine extraction time) was applied. After grinding and separating soluble proteins and fat, the purified raw material was treated in water at pH 7.5 with the addition of endoprotease at 23 ◦C and after thorough washing with water at 80 ◦C, gelatine was extracted. By the suitable choice of process conditions, gelatine with high gel strength (220–320 bloom), low ash content (<2.0%) and viscosity of 3.5–7.3 mPa·s can be prepared. The extraction efficiency was 18–38%. The presented technology is innovative mainly by the enzymatic processing of the source raw material, which is economically, technologically, and environmentally beneficial for manufacturers. Chicken gelatines are a suitable alternative to gelatines made from mammals or fish, and can be used in many food, pharmaceutical, and biomedical applications.

**Keywords:** biotechnology; by-products; chicken feet; extraction; food applications; gelatine; pharmaceutical applications; polymer biomaterials

#### **1. Introduction**

Gelatine is a significant, water-soluble protein that is obtained from collagenous raw materials by partial hydrolysis. The primary raw materials for gelatine production are pork and bovine skins/hides and bones [1]. In recent years, alternative sources of collagen, especially fish and by-products of the meat processing and poultry industry, have become more important for gelatine producers [2]. The reason for this is the growing global demand for gelatine, which is estimated to be around 451,000 tons for 2018: about a 1⁄4 increase over six years [3]. Another impetus for the search for alternative sources of collagen is the growing demand for non-mammalian gelatines, especially from consumers from Islamic, Jewish, and Hindu countries. Also, the economic and ecological reasons for the consideration of by-products are forcing producers of such waste to seek ways of further re-utilization. The new application possibilities of gelatine made from alternative sources of collagen are also opening up.

Worldwide annual food waste production is estimated at up to 100 million tons. In the European Union (EU), the meat processing industry produces 16.5 million tons of waste, and 5.2 million tons of waste is generated in the fish production industry [4]. The United States (US) poultry industry produced 28.4 million tons of live weight of poultry in 2016; the largest proportion (24.9 million tons) was chicken [5]. Poultry meat accounts for 5.5% of total agricultural production and 12.7% of EU meat production. In 2014, 13 million tons of poultry meat was produced in the EU, 78% of which was chicken [6]. According to the FAO (Food and Agricultural Organization), the total consumption of poultry meat is growing by 3.6% per year. Even though in some countries animal by-products (e.g., kidneys, hearts, livers, lungs, stomachs, and tongues) are used for culinary purposes, approximately 20–30% of the live weight of poultry represents unused by-products. In some countries, for example, blood is used as an ingredient in food production and for the production of feeding meal. Feathers and other by-products are used for the production of feeding hydrolysates; part of the protein waste can also be composted or used for fuel production, and waste fats are used for soaps, biofuels, and lubricants [7–10].

Raw materials for the production of commercial gelatines are processed in low and high pH environments [1]. In the case of type A gelatine, the raw material is treated for 18–30 h in an acidic environment at pH 1.5–3.0; this procedure is suitable for pigskins. For type B gelatine, the raw material is treated in alkaline medium at pH 12.0 for several weeks to months; this procedure is mainly used for cowhides.

By-products from freshwater and saltwater fish rich in collagen (skins and bones) have long been known as the suitable raw materials for gelatines and collagen hydrolysates production. In the preparation of fish gelatines, the procedure is analogous to that of the production of gelatines from mammals. Most often, the raw material is treated for 12–48 h in an acidic environment at 5–10 ◦C, optionally at room temperature; weak solutions of acids (0.05–0.5 mol·L<sup>−</sup>1) are used, especially acetic acid, citric acid, phosphoric acid, formic acid, hydrochloric acid, propionic acid, and optionally lactic acid. Depending on the type of the raw material (different amounts of intermolecular and intramolecular crosslinking of the collagen matrix) and the extraction conditions (water, usually at 45–55 ◦C for 3–12 h), gelatine (type A) is prepared with a yield of 5.5–66.0% [11–13]. Less common is the processing of the raw material in an alkaline environment, most often using Ca(OH)2 (type B gelatine), as well as the use of protease [14–16]. Somewhat different is the industrial production of gelatines, soluble collagen, and hydrolysates from poultry tissues. On the global market, chicken products represent only a fraction of gelatines made from beef, pork, and fish. After 24 h of processing of fowl feet with weak acid solutions (0.5 M acetic acid, citric acid, hydrochloric acid, or lactic acid) and enzyme (pepsin) at 4 ◦C, soluble collagen is prepared with relatively low yields, 5.6–8.4% [17]. Using stronger solutions of the same acids (5.0%, v/v), acid-soluble collagen can be prepared under similar processing conditions (12–36 h incubation at 4–7 ◦C) with a yield of 7.9–31.2%, even without the use of pepsin [18]. Also, the preparation of gelatines, especially from chicken feet, skins, and tendons, is described. After the removal of soluble non-collagenous proteins and pigments, the feet are processed in an acidic environment using slightly stronger solutions (1.5–4.5%, v/v) of acetic acid, citric acid, or lactic acid at room temperature for 16–18 h. At moderate extraction temperatures (50–55 ◦C) and extraction times ranging from tens of minutes to hours, gelatine with a gel strength of 120–300 Bloom is prepared, which is comparable to commercial pork and fish gelatines. The disadvantage is the relatively low yields of the prepared gelatines, 6.0–14.5%, based on the dry weight of the source material [19,20]. After combined sulphuric acid and acetic acid treatment and 12-h extraction at 45 ◦C, 16% of high gel-strength gelatine (355 Bloom) can be obtained from chicken skins, which is more than in conventional commercial bovine gelatine [21]. The literature also describes the processing of other poultry by-products into gelatines. These are mainly chicken or turkey heads or duck feet. The processing is carried out according to an analogous procedure to that of chicken feet or skins. The alkaline method of poultry tissue processing is not widespread, but has been tried, for example, on chicken and turkey heads [22].

The processing of some by-products from poultry slaughter into gelatines have been satisfactorily described in the literature. However, all the techniques are based on the acidic or alkaline processing of the raw material; the enzymatic processing is only marginally mentioned. The use of enzymes

brings with it many advantages, such as mild reaction conditions given by temperature and pH as well as low doses of enzymes used. The main goal of our paper is to propose a biotechnological method of preparing gelatines from chicken feet, through first treating the source material with a suitable proteolytic enzyme and then extracting the gelatine with hot water. This is followed by observing the influence of selected process factors on gelatine yield (percentage conversion of the source material to gelatines) and the quality of prepared gelatines (gel strength, viscosity, and ash content in gelatine). The specific hypothesis being tested is that chicken feet, after treatment with endoprotease and extraction with hot water, can be processed into gelatines whose properties are comparable with gelatines produced from pork and beef tissues.

#### **2. Materials and Methods**

A flow chart of biotechnological processing of chicken feet into gelatines is shown in Scheme 1.

**Scheme 1.** A flow chart of preparation of gelatins from chicken feet.

#### *2.1. Materials*

Chicken feet were obtained in the Raciola poultry farm (Ltd., Uherský Brod, Czech Republic) and processed under strict hygiene conditions, see Section 2.3 and Section 2.4. Raw chicken feet composition: 35.0 ± 3.0% dry matter content; dry matter: 48.3 ± 0.4% protein content (82.8 ± 0.7% collagen content), 34.8 ± 0.8% fat content, 16.1 ± 0.2% ash content. The polarzyme 6.0 T-granulated endoprotease that hydrolyzes internal peptide bonds were manufactured by the fermentation of a microorganism that is not present in the final product (Novozymes, Copenhagen, Denmark), petrolether, ethanol, NaOH, 36% HCl. The procedure also used a SPAR Mixer SP-100AD-B industrial meat cutting machine with a four-arm knife (Panther T & H Industry, Taichung, Taiwan), Nabertherm muffle furnace (Lilienthal, Germany), WTW pH 526 pH meter (WTW, Oberbayern, Germany), Kavalier LT3 shaker (Sázava, Czech Republic), Kern 770 analytical and precision balances (Balingen, Germany), Kern 440-47 laboratory scales (Kern, Balingen, Germany), MEMMERT ULP 400 dryer (Schwabach, Germany), WTB Binder

E-28-TB1 dryer (Tuttlingen, Germany), IKA LABORTECHNIK RCT BASIC magnetic mixer with heating plate (Staufen, Germany), stainless steel sieve with 1.0-mm size, Stevens-LFRA Texture Analyser (Brookfield, UK).

#### *2.2. Strategy*

A two-stage biotechnological procedure was designed for the conversion of collagen to gelatines, which due to its variability of the process parameters would allow preparing the gelatines of the desired properties with the optimum utilization of the source material. In industrial practice, factor experiments are widely used to design experiments consisting of multiple technological variables, allowing maximum information to be obtained and optimal process conditions to be designed [23].

To study the influence of selected process factors on gelatine yield and the quality of prepared gelatines, 2<sup>3</sup> factor schemes with one central experiment and one repetition were used; then, in the optimization part of the process, 22 factor schemes with one central experiment and one repetition were used. The process factors studied were: factor A—enzyme addition: a minimum value of 0.2% (based on the protein dry matter, w/w), a mean value of 0.5% (w/w), a maximum value of 0.8% (w/w); factor B—enzyme treatment time: minimum 24 h, mean 72 h, maximum value 120 h; factor C—gelatine extraction time: minimum 1 h, mean 2.5 h, maximum value 4 h. The variables evaluated were gelatine yield (percentage conversion of the starting protein substrate to gelatine), the strength of the gelatine gels, the gelatine viscosity, and the ash content of the gelatines. On the basis of statistical evaluation of the influence of process parameters on the evaluated quantities, the process optimization was carried out. The process was based on monitoring the influence of the two most important process parameters on the evaluated variables and the design of optimal processing conditions for the conversion of collagen protein of chicken feet to high-quality gelatines.

#### *2.3. Grinding and Preparation of Raw Material for Biotechnological Processing*

When preparing the starting material for further biotechnological processing, it is necessary to set conditions to avoid the denaturation of the collagen protein, convert collagen to gelatine in biotechnological processing with optimum efficiency, and produce quality gelatines. Immediately after separating the feet, the raw material was rinsed in water and cooled to 0–5 ◦C to prevent negative microbial growth (the chilled raw material can be stored for a maximum of 36 h after slaughter).

Prior to two-phase grinding (on a meat cutter), the feedstock was slightly pre-frozen at −2 to −5 ◦C in the core of the feedstock. In the first milling process, kidney-shaped cutting plates with a hole size of 20.0–30.0 mm were used, the second was a hole size of 3.0 mm; during grinding, the temperature of the raw material increased to a maximum of 3 ◦C. If the ground and homogenized raw material is not processed within a maximum of 24 h, it must be wrapped in vacuum containers with a minimum wall thickness of 80 μm and deep-frozen at −36 ± 2 ◦C; the frozen raw material can be stored at −20 ± 2 ◦C for up to 24 months. Furthermore, the homogenized raw material needs to be purified from the accompanying non-collagenous components, pigments, and fat. The removal of albumins, globulins, and pigments was based on the earlier procedure [24] with slight modifications. The raw material was mixed with 0.10% NaOH solution in a ratio of 1:8 (w/v) and shaken for 45 min at room temperature; then, the solution was filtered through a 1.0-mm stainless steel sieve, and the raw material was washed with running water; the whole procedure was repeated four times. Then, the raw material was dried in an air circulating oven at 35.0 ± 0.5 ◦C (36 h).

The defatting of poultry tissues is described in the literature by several methods, such as the Soxhlet extraction process [21], using NaHCO3 [24] or by mechanical separation after fat leaching in alkaline medium [25]. Defatting of the raw material was carried out according to the procedure previously described [26]. The dried raw material was mixed with a mixture of ethanol and petroleum ether (1:1, v/v) solvent in a ratio of 1:6 (w/v) and shaken for 32 h at room temperature; after 8 h, the solvent mixture was replaced with a new one (filtration was carried out through a stainless sieve with a size of 1.0 mm). The defatted raw material was spread on a metal sheet, and the solvent residues were evaporated in a fume hood at room temperature.

#### *2.4. Biotechnological Processing of Raw Material and Gelatine Extraction*

As there is a lack of biotechnological processing of the raw material prior to gelatine extraction in the available literature, such a procedure has been designed and tested in our workplace. We proceeded from the results of our previous research, which focused on the preparation of protein hydrolysates from chicken feet [27]. The defatted raw material was mixed with distilled water in a ratio of 1:10 (w/v); after 30 min, the pH of the mixture was adjusted to 7.5 ± 0.3 (optimal efficiency of the proteolytic enzyme used) by the addition of 10% HCl. After the addition of Polarzyme 6.0 T in an amount according to factor A (enzyme load based on the dry weight of the defatted raw material), the mixture was shaken at room temperature for a time according to factor B. After filtration (through a 1.0-mm stainless steel sieve), the protein hydrolysate solution was first brought to boil and was boiled for 10 min (inactivation of proteolytic enzyme). Then, it was dried (after pouring onto a stainless steel plate) in a thin film in an air-circulating oven at 45.0 ± 1.0 ◦C (48 h); afterwards, the dried film was scraped off and weighed. The enzyme treated raw material was thoroughly washed on the sieve under running water and then mixed in a beaker with distilled water in a ratio of 1:9 (w/v). The mixture was placed on a hot plate, stirred gently (magnetic stirrer), heated to 80.0 ± 0.5 ◦C (d*t*/dτ = 10 ◦C/min), and after reaching this temperature, gelatine was extracted according to the factor C, time. After completion of the extraction, the gelatine solution was separated by filtration through a 1.0-mm stainless steel sieve equipped with three layers of polyamide fabric (300-μm pore size). The gelatine solution was brought to the boil and was boiled for 5 min; then, it was poured onto a thin-plate thin film and dried in an air-circulating oven at 45.0 ± 1.0 ◦C (48 h); afterwards, the dried film was scraped off and weighed. The undissolved residue of the raw material after gelatine extraction was dried in an air circulating oven at 103.0 ± 1.0 ◦C (for 16 h) and then weighed.

#### *2.5. Analytical Methods*

Dry matter, ash, fat, and protein were determined by conventional food methods [28–30]. The dry matter was determined by the indirect method of drying the sample for 18 h at 103.0 ± 2.0 ◦C; the ash was determined gravimetrically after burning and annealing the sample; fat was determined by Soxhlet extraction; nitrogen was determined by the Kjeldahl method, and the protein content was calculated from the determined nitrogen content by multiplying by a factor of 6.25. Collagen content was calculated from the hydroxyproline content (determined colorimetrically after sample hydrolysis in 6 mol·L−<sup>1</sup> HCl) by multiplying by a factor of eight [31,32]. Gel strength, gelatine viscosity, and pH were determined according to the Official Procedure of the Gelatine Manufacturers Institute of America [33]. The pH of a 1.5% gelatine solution was determined by potentiometry at a temperature of 35 ± 0.5 ◦C using a pH meter. The gelatine gel strength was determined from a gel formed from a 6.67% solution prepared according to prescribed conditions by the measuring of force (weight) required to depress a prescribed area of the surface of the sample to a distance of 4 mm. The dynamic viscosity of a 6.67% gelatine solution was determined at 60 ◦C by measuring the flow time of 100 mL of the solution through a standard pipette; the viscosity was calculated from Equation (1). The yield of the hydrolysate was calculated from the weight of the hydrolysate prepared after the biotechnological treatment of the raw material, the yield of gelatine from the weight of the gelatine (both yields based on the weight of the defatted raw material); further, the total yield was calculated; see Equations (2, 3, 4). The mass balance error is expressed by the percentage difference of the dry matter mass balance between the input (defatted raw material) and the outputs (hydrolysate, gelatine, and undissolved residue); see Equation (5).

$$
\ln \eta = \left(\mathbf{A}\boldsymbol{\pi} - \frac{\mathbf{B}}{\boldsymbol{\pi}}\right) \mathbf{d} \tag{1}
$$

$$\text{HY} = \frac{\text{m}\_1}{\text{m}\_0} \, 100 \,\tag{2}$$

$$\text{GY} = \frac{\text{m}\_2}{\text{m}\_0} \, 100 \, (3) \tag{3}$$

$$\mathbf{Y}\_{\Sigma} = \mathbf{H}\mathbf{Y} + \mathbf{G}\mathbf{Y} \tag{4}$$

$$\text{MBE} = \frac{\left\lfloor \left( \text{m}\_1 + \text{m}\_2 + \text{m}\_3 \right) - \text{m}\_0 \right\rfloor}{\text{m}\_0} 100 \tag{5}$$

where η is gelatine viscosity (mPa·s), *A* and *B* are pipette constants, τ is efflux time (s), *d* is solution density (for a 6.67% gelatine solution at 60 ◦C d = 1.003), *HY* is the hydrolysate yield (%), *m0* is the weight of the defatted raw material (g), *m1* is the weight of the hydrolysate, *GY* is the gelatine yield (%), *m2* is the weight of gelatine (g), *m3* is the weight of the undissolved residue (g), Y is the total yield (%), and *MBE* is a mass balance error (%).

#### *2.6. Statistical Analysis*

A two-level factorial design of experiments and evaluation of the results were carried out with Minitab® 17.2.1 software (Fujitsu Ltd., Tokyo, Japan). Statistical analyses (arithmetic means and standard deviations) were accompanied using Excel 2010 (Microsoft, Inc., Seattle, WA, USA) at the significance level of 5% (*P* < 0.05).

#### **3. Results and Discussion**

#### *3.1. Study of the Influence of Process Factors on the Gelatine Yield and Quality of Prepared Products*

A schedule of the experiments and summary results of the processing of chicken feet proteins into gelatine and hydrolysates by two-level factor schemes with three factors of concern are given in Table 1. The pH of the prepared gelatines ranged from 6.0 to 6.4, which corresponds to standards for food and pharmaceutical gelatines where pH 4.0–7.5 is prescribed. The pH of commercially produced porcine gelatines ranges from 5.5 to 6.5, with beef gelatines usually between 5.5–7.0, as well as fish gelatines. All the prepared gelatines were characterized by very low ash content (0.61–1.66%), and thus meet the stringent parameters for food and pharmaceutical gelatines (Food Chemical Codex 10; United States Pharmacopoeia 35 NF 30; European Pharmacopoeia; Japanese Pharmacopoeia 15).

**Exp. No. Factors under Study Collagen Hydrolysate Gelatine Summary of ABC the Process Addition (%) Treatment (h) Time (h) HY (%)** ± **SD (%) GY (%)** ± **SD (%) SD (Bloom) SD (mPa**·**s) pH**± **SD Y(%) MBE (%)**



a—based on dry matter; HY—hydrolysate yield; GY—gelatine yield; Y—total yield; AshH—ash content in hydrolysate; AshG—ash content in gelatine; F—gelatine gel strength; η—gelatine viscosity; MBE—mass balance error

The statistical significance of the studied process factors in the observed limits was evaluated using the standard Fisher's significance test and the P-values for a 95% confidence level. For the factor schemes used by us, the critical value is *F* = 10.13, so the higher the *F*-value is above the critical value, the greater the influence of the process factor. Similarly, results for *P*-values were evaluated; factors with a value lower than α = 0.05 have an effect on the evaluated variables with 95% probability, and the lower the *P*-value, the greater the influence of the process factor [34]. Table 2 shows the results of analysis of variance for gelatine yield, gelatine gel strength, and gelatine viscosity.


**Table 2.** Analysis of variance of the experimental design for gelatine yield, gelatine gel strength, and gelatine viscosity (study of the influence of process factors).

#### 3.1.1. Gelatine Yield

The results of the statistical evaluation showed that only factor A (enzyme addition) is statistically significant for the yield of gelatine (GY). The effect of the two most important process factors (enzyme addition and gelatine extraction time) on GY is represented by the contour graph in Figure 1. It is evident that with increasing enzyme addition and at the same time extending the extraction time, GY increases. This trend is particularly evident at lower enzyme additions (up to about 0.4%) and at shorter extraction times (up to about 2.5 h). There is no significant GY growth with any addition of enzymes above 0.5% or extraction time more than 2.5 h. The minimum GY (≈ 21%) is reached under the lower limit of the factors of interest, i.e., 0.2% enzyme addition and 1.0-h extraction time. The maximum GY (≈ 38%) then corresponds to approximately 0.7% enzyme addition and 2.5 h; further increasing the extraction time will no longer affect the increase of GY.

GY values are comparable or better compared to the available results of processing the proteinic poultry tissues into gelatines and hydrolysates. Using an acid extraction procedure and ultrasonic extraction, very low yields of gelatines (4% and 17%) from chicken feet are reported [35]. Almeida, Calarge, and Santana processed chicken feet by pressure extraction in water (120 ◦C for 20 min) and obtained a 36% yield of low-strength gelatine [36]. Du et al., by processing the chicken and turkey heads in an acidic environment and extracting in two stages (temperatures 50 and 60 ◦C), prepared high-quality gelatines with yields of 21% and 31% (chicken heads) and 25% and 38% (turkey heads) [24]. Sarbon, Badii, and Howell used the combined alkaline-enzyme processing of raw material to produce high-quality chicken skin gelatine, with a gelatine yield of 16% [21].

**Figure 1.** Effect of enzyme addition and gelatine extraction time on gelatine yield (study of the influence of process factors).

#### 3.1.2. Gelatine Gel Strength

The results of the statistical evaluation showed that all three studied factors are statistically significant in the strength of gelatine gel (*F*); the influence of factor A (enzyme addition) and factor B (enzyme treatment time) on F is represented by the contour graph in Figure 2. It can be seen from the figure that high-quality gelatines (*F* = 220–280 Bloom) can be prepared within the limits of the process factors studied (0.2–0.8% enzyme addition and raw material enzyme processing for 30 to 120 h). Under the lower limit of the observed factors (0.2% enzyme addition and 30-h enzyme treatment of the raw material, gelatine with *F* > 280 Bloom can be prepared with a 21% yield of gelatine (see Figure 1). With an almost double yield of gelatine (GY = 38%, see Figure 1) under the upper limits of the observed factors (0.8% enzyme addition and 120-h enzyme treatment of the raw material), it is evident that there is no significant decrease in F: gelatine still has a high gel strength (F ≈ 220 Bloom). The standard specifications for food gelatines prescribe *F* = 150–280 Bloom, depending on the method of application. For the production of hard gelatine capsules (HGC), *F* = 200–280 Bloom is prescribed; gelatines with a gel strength of 130–200 Bloom are sufficient for soft gelatine capsules (SGC).

**Figure 2.** Effect of enzyme addition and enzyme treatment time on gelatine gel strength (study of the influence of process factors).

The gel strength of gelatines prepared under various process conditions is the same or higher than that of commercial high-strength gelatines (*F* > 200 Bloom) made from bovine and pork hides/skins and bones and fish. The gelatines prepared by us have a comparable or lower F compared to gelatines prepared from poultry by-products by other authors. Gelatine prepared from chicken feet by acid treatment of the starting material had a gel strength of 295 Bloom [20]. Du et al. reported *F* = 200–248 Bloom in gelatines prepared from chicken heads; however, higher F values (333–368 Bloom) were achieved in gelatines prepared from turkey heads [24]. The results are similar in high-quality gelatine (*F* = 355 Bloom) from chicken skins [21].

#### 3.1.3. Gelatine Viscosity

The results of statistical evaluation showed that factor A (enzyme addition) and factor B (enzyme treatment time) are statistically significant for gelatine viscosity (η); the influence of these process factors on η is represented by the contour graph in Figure 3. It is evident from the figure that η is the highest (>6.5 mPa·s), and at the same time, its value is not influenced by the change of these technological factors at low enzyme addition (0.2–0.3%) and at short enzymatic treatment times (within 90 h). It is also evident from the graph that with increasing enzyme addition (>0.4%) and at short enzyme treatment times (up to ≈ 90 h), η gradually decreases to 5.0 mPa·s. Long enzyme treatment times (110–120 h) in combination with the highest enzyme additions (0.7–0.8%) mean an even more significant decrease of η (3.5–4.0 mPa·s). Gelatines with η 2.0–7.5 mPa·s cover a wide range of applications in the food industry. For the production of HGC, gelatines with a higher viscosity (4.5–6.5 mPa·s) are required. For SGC, it is 2.5–4.5 mPa·s, and for tablets, the viscosity of 1.7–3.5 mPa·s is sufficient. It can be seen from the viscosity measurement results that gelatines prepared under different process conditions meet the gelatine specifications for the production of HGC and SGC.

**Figure 3.** Effect of enzyme addition and enzyme treatment time on the gelatine viscosity (study of the influence of process factors).

#### *3.2. Process Optimization*

From the study of the influence of process parameters on the gelatine yields and the quality of the prepared gelatines, the following conclusions can be drawn: *a)* with increasing enzyme addition (factor A) and with increasing gelatine extraction time (factor C), the gelatine yield increases; *b)* with increasing enzyme addition (factor A) and with increasing enzyme treatment time (factor B), the strength of the gelatine gels decreases; *c)* the highest gelatine gel strength (≈ 295 Bloom) was recorded under the conditions of the lower process factors observed (0.2% enzyme addition, 24-h enzyme treatment time and 1-h gelatine extraction time); *d)* it can be assumed that the strength of the gelatine gels will grow at lower enzyme additions, shorter enzyme treatment times, and shorter extraction times.

When optimizing the process conditions for the preparation of gelatines from chicken feet, the goal was to prepare high-quality gelatines (with a gel strength of at least 300 Bloom) even at the cost of a lower gelatine yield. It was decided to enzymatically process the raw material for a constant time (20 h) and monitor the effect of enzyme addition (0.1–0.4%) and gelatine extraction time (15–45 min) on the quality of the prepared gelatines (gel strength, viscosity, ash content) and gelatine yield. A schedule of experiments (factor schemes 22) and summary results of the optimization part of processing chicken feet proteins to gelatines are given in Table 3.

**Table 3.** The experimental design and the results of processing of chicken feet into gelatines (process optimization).


a—based on dry matter; GY—gelatine yield; AshG—ash content in gelatine; F—gelatine gel strength; η—gelatine viscosity

As shown in Figure 4, gelatines with a very high gel strength can be prepared by the appropriate choice of enzyme addition (factor A) and gelatine extraction time (factor C). In the case of low enzyme addition (up to 0.15%) and short extraction times (up to 20 min), the highest quality gelatine (F = 320–325 Bloom) is prepared; the gelatine yield is approximately 17.5–18.0% (see Figure 5). By increasing the enzyme addition to the upper test limit (0.4%) and extending the extraction time (up to 45 min), the gelatine yield increases to about 21.0% (Figure 5), but gelatine still has a very high gel strength (F ≈ 305 Bloom); see Figure 4.

**Figure 4.** Effect of enzyme addition and gelatine extraction time on gelatine gel strength (process optimization).

**Figure 5.** Effect of enzyme addition and gelatine extraction time on gelatine yield (process optimization).

The pH of the prepared gelatines ranged between 6.9–7.3 and the ash content was 1.23–1.94%, which meet both the food standards where the permitted ash content is up to 2.0% (Food Chemical Codex 10) and pharmaceutical standards where the permitted ash content is up to 3.0% (United States Pharmacopoeia 35 NF 30; European Pharmacopoeia; Japanese Pharmacopoeia 15). Gelatines have a higher viscosity (6.8–7.3 mPa·s) and are suitable, for example, in the food industry for the production of deposited marshmallows, or as a binder and gelling agent for meat industry products or for coatings.

The observed technological conditions in the preparation of gelatines have an analogous effect on the yield of gelatines and their quality, as is the case with the production of gelatines from traditional raw materials (beef and pork hides/skins and bones, and fish) by established technologies, i.e., by processing the raw material in an acidic or alkaline environment [1]. By properly selecting the amount and time of action of the enzyme in the processing of the raw material and the appropriate extraction time, gelatines with the desired gel strength and viscosity can be prepared. In industrial practice, it will be advantageous to utilize the multiple fractionation of the raw material at increasing extraction temperatures (typically 60–100 ◦C), which will result in the preparation of different quality gelatines with optimum utilization of the starting material (maximal gelatine yield). The undissolved residue after the gelatine extraction can be processed by hydrolysis with a suitable proteolytic enzyme into a collagen hydrolysate, which can be used in the food industry (e.g., as nutritional supplements, thickeners, etc.) or as an additive (humectant) in the cosmetic industry.

From the study of the influence of process factors and from the optimization part of the biotechnological preparation of gelatines from chicken feet, it can be stated that gelatines with a gel strength of 220–320 Bloom, viscosity of 3.5–7.3 mPa·s, and an ash content of up to 2.0% can be prepared by a suitable combination of process factors (amount of enzyme, time of raw material enzyme treatment, time and temperature of gelatine extraction). These are gelatines belonging to the category of high-quality food and pharmaceutical gelatines (gel strength of 220–300 Bloom, viscosity of 2.0–7.5 mPa·s) complying with the strict standards prescribed in the European Union, US or Japan (Directorate General for Health and Consumer Protection of the European Commission, USDA, FDA or MHLW).

Chicken gelatines with a high gel strength and a viscosity of about 4.0–5.5 mPa·s are suitable, for example, for the production of gelatine desserts, confectionery products (e.g., gummy bears, extruded marshmallows), and are also perfect for use in the meat industry (aspics, binder for meat emulsions, ham, jellies), in the manufacture of dairy products (low-fat butter spreads, panna cotta, aerated desserts, yogurt-based products), or on frozen semi-finished or finished products. Conversely, gelatines with a lower viscosity (<4.0 mPa·s) are suitable for making chewing gums or sugar sanded/oiled

jelly items. For the preparation of hard gelatine capsules, all the prepared chicken gelatines with a viscosity of 4.5–6.5 mPa·s are suitable; for producing soft gelatine capsules, gelatines with a viscosity of 2.5–4.5 mPa·s are sufficient. Due to the high gel strength, chicken gelatines may find various applications in the biomedical field, similar to gelatines produced from pork and bovine feedstock. For example, they can be used as hydrogel carriers for delivering bioactive macromolecules, producing membranes, microspheres, and nanoparticles, encapsulating carriers for the controlled release of biologically active substances, or delivering cell transplants for tissue repairs [37–39]. A high Bloom value has a positive effect on the mechanical properties of gelatine films [40]. In addition, Lai suggested that the gelatine Bloom value has an influence on cellular responses to gelatine material as well [41]. The Bloom value of chicken gelatines is a critical parameter for preparing mixtures with other biopolymers, above all with polysaccharides, to adjust the gelling/thickening properties of such mixtures for food and pharmaceutical applications [42].

#### **4. Conclusions**

Most commercial gelatines are made from beef, pork, or fish tissues; industrial processing is based on the acidic or alkaline processing of the feedstock. The presented technology focuses on the preparation of gelatines from by-product collagen raw materials derived from the slaughter of chicken (chicken feet). The innovative technology element brings the biotechnological processing of (purified) feedstock by commercial food endoprotease, which, in contrast to acidic (type A gelatines) or alkaline (type B gelatines) processing, has a variety of economic, technological, and environmental advantages. The feedstock is treated with a small amount of enzyme (0.1–0.8%, based on the dry raw material weight) in neutral environment at room temperature; and the processing time, e.g., compared to alkaline beef processing, is significantly reduced (from weeks to months to tens of hours). After enzymatic treatment and washing of the feedstock in water, a standard hot-water extraction of gelatine follows. By selecting the technological conditions, gelatines with high gel strength (220–320 Bloom) with an ash content of less than 2.0% can be prepared to meet food and pharmaceutical standards. The gelatine yield is 18–38% for the one-step extraction process, which are very good values; a more optimal utilization of the feedstock can then be achieved in practice by a multi-step extraction process. The observed process parameters in the preparation of chicken gelatines also affects the gelatine viscosity (3.5–7.3 mPa·s). The strength of gelatine gels (a key parameter for the application of gelatines) is not significantly altered by the change in process parameters, as all of the gelatines prepared belong to the gelatine category with a high gel strength (>220 Bloom). Chicken gelatines can be an alternative to beef, pork, and fish gelatines, as they meet halal and kosher requirements. For their excellent properties, they are suitable in many food applications; for example, for confectioneries, sweet desserts, dairy products, or meat products. They can also be used in the pharmaceutical field for manufacturing hard gelatine capsules (HGC) and soft gelatine capsules (SGC), as well as various applications in the biomedical field (hydrogels, membranes, carriers, and films).

#### **5. Patents**

From the work reported in this manuscript, the following patent resulted: Patent CZ 307665 – biotechnology-based production of food gelatine from poultry by-products.

**Author Contributions:** Supervision, visualization, writing—original draft preparation, writing—review and editing: P.M. (Pavel Mokrejš); methodology, validation: P.M. (Pavel Mokrejš) and P.M. (Petr Mrázek;) formal analysis, resources: P.M. (Petr Mrázek;) funding acquisition, project administration: R.G.; software: J.P.; data curation, investigation: P.M. (Petr Mrázek) and R.G.

**Funding:** This research was funded by the Internal Grant Agency of the Faculty of Technology, Tomas Bata University in Zlín, ref. IGA/FT/2019/003.

**Acknowledgments:** The authors thank to David Dohnal (Olomouc, the Czech Republic) for professional editing of the manuscript.

**Conflicts of Interest:** The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.

#### **References**


© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
